Technology policy for the knowledge economy: Public support to young ICT service firms

ARTICLE IN PRESS

Telecommunications Policy 31 (2007) 573–591 www.elsevierbusinessandmanagement.com/locate/telpol

Technology policy for the knowledge economy: Public support to young ICT service firms Massimo G. Colombo, Luca Grilli Politecnico di Milano, Department of Management, Economics, and Industrial Engineering, P.za Leonardo da Vinci 32, 20133 Milan, Italy

Abstract Public intervention in high-tech sectors is often advocated to resolve market imperfections that may possibly limit the viability of young high-tech enterprises. Although some European countries have adopted national government support policies that explicitly target this type of firm, in Italy and in other EU countries, there are no national support measures specifically designed for them. The paper focuses on the information and communication technologies (ICT) services sector in Italy. It aims to investigate whether both horizontal general-purpose direct support mechanisms at the national level and financial support measures provided by local administrative entities (which rarely have been specific to the ICT sector) permit an efficient allocation of public funds. r 2007 Elsevier Ltd. All rights reserved. Keywords: Young ICT service firms; Horizontal direct public subsidies; Italy

1. Introduction The development of information and communication technologies (ICTs) plays a key role in contributing to improved productivity of labor and capital and in fostering economic growth, especially in advanced economies (OECD, 2002, 2004; Schreyer, 2000). High-tech startups in ICT-related activities have represented the fundamental engine of this process (Acs, 2004), and many young US firms (e.g., Microsoft, Cisco, Yahoo!, Google, Amazon) have attained worldwide leadership in the ICT sector, in both the manufacturing and service sectors, in a short period of time. Even if ICTs are defined as general-purpose technologies, there is growing evidence that the emergence and consolidation of a strong national ICT sector are a fundamental prerequisite for rapid ‘‘digitalization’’ of a country (Guillen & Suarez, 2001) and have a positive impact on the performance of the whole national economy (Daveri & Silva, 2004; Gordon, 2000). Europe has traditionally lagged behind the United States both in the creation of ICT startups and in the number and size of young high-growth firms in this sector (the so-called ‘‘gazelles’’).1 The favorable ecosystem in terms of innovation, entrepreneurial culture, functioning of capital and labor markets, and level of human capital in which new US Corresponding author. Tel.: +39 02 2399 3955; fax: +39 02 2399 2710.

E-mail address: [email protected] (L. Grilli). The only young European ICT service firm to have achieved results comparable to those achieved by American high-tech startups is the German SAP firm, a worldwide leader in the production of enterprise resource planning (ERP) software. 1

0308-5961/$ - see front matter r 2007 Elsevier Ltd. All rights reserved. doi:10.1016/j.telpol.2007.08.001

ARTICLE IN PRESS 574

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

high-tech ventures grow to maturity certainly contributes to explaining this difference in performance. As a result, public intervention to favor high-tech entrepreneurship is often advocated in Europe to overcome possible market imperfections and to sustain the creation and development of new ICT ventures that can compete in the global arena. Basically, economic analysis has developed two main rationales for public sector support to new technology-based firms. First, the socially optimal level of R&D expenditures may be higher than the optimal level for private individuals because of the presence of R&D spillovers. Young, small firms may invest less than the social optimum because they are unable to defend innovation and extract most of the rents in the product market (Griliches, 1992; Jaffe, 1996; Teece, 1986). Secondly, a large body of empirical literature on entrepreneurship has pointed to the presence of financial constraints on new firms (Black, de Meza, & Jeffreys, 1996; Blanchflower & Oswald, 1998; Evans & Jovanovic, 1989; Evans & Leighton, 1989; Holtz-Eakin, Joulfaian, & Rosen, 1994a, b; Meyer, 1990). Access to the credit market is considered to be problematic, especially for high-tech ventures (Carpenter & Petersen, 2002; Storey & Tether, 1998; Westhead & Storey, 1997; see also Colombo & Grilli, 2007 and Grilli, 2005 for the Italian context). Many obstacles to external financing for this type of enterprise stem from the inability of banks and other financial institutions to distinguish good projects from ‘‘lemons’’ in sectors usually characterized by highly skewed returns, asymmetric information, both ex-ante and ex-post (e.g., hidden information and hidden actions), and a lack of inside collateral to secure debt (Carpenter & Petersen, 2002). On the other hand, even though for high-tech startups private equity financing has advantages over debt (Carpenter & Petersen, 2002), this mode of financing may still present problems related to ex-ante asymmetric information (Myers & Majluf, 1984) and high transaction costs (Asquith & Mullins, 1986; Lee, Lochhead, Ritter, & Zhao, 1996) that inhibit access to seed and startup equity capital for most new high-tech ventures. Both arguments lead to the same conclusion: valuable innovative projects may be disregarded and remain unrealized because of spillover problems or lack of sufficient funds. Naturally, from a social point of view, these are missed opportunities calling for public intervention. But which form of public intervention is most suitable for sustaining young ICT service firms while avoiding distortions,2 reducing the risk of wasting public resources, and consequently maximizing social well-being? Policymakers have a wide spectrum of measures at their disposal. First, it is important to understand whether young ICT service firms need specific, customized programs (i.e., a vertical technology policy) or whether they may be effectively supported through horizontal programs with more general objectives (e.g., a program targeting support to innovation or entrepreneurship). Secondly, the question also arises whether national centralized governmental bodies or local ones are best suited to provide public support to this type of firm. Thirdly, it is also important to understand whether direct assistance programs or indirect ones (i.e., support to institutions that provide financing and other services to new high-tech ventures) are more efficient. As was recently suggested by Siegel, Wessner, Binks, and Lockett (2003), assessment of the efficiency of public policy measures designed to promote innovation in high-tech firms and to resolve market inefficiencies has become a key policy issue. If evaluation methodologies present many issues and are a challenging field of research (see European Commission Joint Research Centre, 2002 for a comprehensive treatment of the issue, and also Lerner, 1999, 2001), there is less disagreement on the characteristics which a program should possess to be truly efficient. First, the program should avoid generating substitution effects in the target sector (Santarelli & Vivarelli, 2002). These occur when inefficient subsidized firms, because of the public aid they receive, have an artificial advantage over potentially more efficient but nonsubsidized ones. Under these circumstances, public intervention may end up distorting market process dynamics, hindering selection, and even promoting artificial incumbency of inefficient firms to the detriment of potentially more efficient competitors and new entrants. Secondly, the program should limit the risk of directing subsidies to 2 Several contributions have emphasized the distortions that may result from public subsidies. As articulated by Holtz-Eakin (2000) and Santarelli and Vivarelli (2002), failure rates are naturally high among new, small firms, and public support may only disturb and delay the competitive selection process, subsidizing inefficiencies. First, screening difficulties may be magnified for policymakers, and second, politicians may seek to direct subsidies in a manner that creates political consensus and benefits themselves rather than citizens (see Lerner, 2002 and the literature mentioned there).

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

575

beneficiaries that do not need public aid to realize their potential. If this does occur, then the policy measure does not produce any positive additional results with respect to the prior situation and in fact generates only deadweight effects (Santarelli & Vivarelli, 2002). According to this view, policymakers should preferentially channel public subsidies to firms that: (i) exhibit genetic characteristics that distinguish high-growth-prospect firms from low performers, and (ii) are financially constrained and because of capital market imperfections cannot obtain the financing they need to realize their growth potential.3 Clearly both conditions are necessary for a policy measure to be efficient, and if one of these two prerequisites is absent, this should lead to a reconsideration of the public policy intervention as implemented. Although most existing research studies evaluate the effectiveness of programs mainly by comparing the performance of subsidized and nonsubsidized firms (see Lerner, 1999, 2002; Lerner & Kegler, 2001 for contributions that refer to new programs targeted at high-tech ventures), in the authors’ view, it is necessary to take a step back and examine which characteristics enable firms to obtain access to direct public financial subsidies. More specifically, the aim of this paper is to focus on condition (i) and analyze empirically whether direct general-purpose support mechanisms at the national level and financial support measures provided by local administrations (which have been predominantly horizontal in nature) are compatible with a scenario where public funds are optimally allocated to young ICT service firms. In fact, if condition (i) is not met, this is a sufficient condition for a policy measure to be inefficient. In this context, the Italian experience is very interesting. In fact, Italy, like many other EU countries, has never had a national financial support program exclusively targeted to young high-tech firms in the ICT or any other sector (Storey & Tether, 1998), but these firms have benefited extensively from public assistance through measures that were also available to broader groups of firms. Furthermore, there has been no systematic indirect support to young ICT service ventures. In consequence, it is legitimate to question whether direct general-purpose policy measures have been efficient in supporting young ICT service firms. If it were determined that they are not efficient, this would call for a new technology policy approach towards this kind of firm and likely would require the implementation of more specific and customized measures (i.e., a vertical approach) rather than the current programs. The paper takes advantage of a new data set based on a sample of 351 young Italian firms operating in the ICT service industry. Condition (i) can be tested by analyzing whether the genetic characteristics that are typical of young, high-growth prospect ICT service firms can explain the access of these firms to public subsidies at both the national and local levels. Data were provided by the Research on Entrepreneurship in Advanced Technologies (RITA) database developed at Politecnico di Milano. The firms in the sample were established in 1980 or later, were independent at the time they were founded and were still independent on 1 January 2004 (i.e., they were not controlled by another business organization, even though other organizations may hold minority shares). Highly detailed corporate-level data are available on each firm’s activities, structure, and performance, on the characteristics of the founding team such as education level and prior working experience, and on access by each firm to national and local public direct support measures. Clearly the sample is very heterogeneous with respect to the variables of interest and in actual firm growth performance. The paper proceeds as follows. Section 2 reviews the extant empirical literature on the determinants of highgrowth performance in high-tech sectors to identify the genetic characteristics that are usually associated with best performers in these markets. Section 3 presents the data set, Section 4 provides some descriptive statistics which depict the ability of firms to access public direct support measures (both at the national and local levels) and highlights possible basic inefficiencies in the allocation of funds. Section 5 is devoted to an econometric exercise that tests whether public funds have been directed preferentially to those young ICT service firms whose genetic characteristics are typical of ‘‘gazelles’’ in high-tech sectors. The test will enable an evaluation of whether or not existing direct public subsidies for young ICT service firms are likely to have generated important substitution effects in the sector. A final section with a discussion of the main findings and some policy implications concludes the paper. 3 It should also be noted that previous studies suggest that firms characterized by greater entrepreneurial ability are the most likely to suffer from financial constraints (see A˚stebro & Bernhardt, 1999; Colombo & Grilli, 2005a; Grilli, 2005). This means that the greater the extent to which a firm has the (i) characteristic, the more likely it is to be in the (ii) condition.

ARTICLE IN PRESS 576

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

2. Young high-growth prospect ICT service firms Any technology policy measure intended to help young firms and maximize social well-being should satisfy two conditions (Santarelli & Vivarelli, 2002). First, it must target those firms that are potentially successful and are characterized by high-growth prospects. If this is the case, the risk of generating a substitution effect is alleviated. This latter term describes a situation in which inefficient firms that manage to obtain a subsidy crowd out (in terms of survival and growth performance) potentially efficient but nonsubsidized firms. The second prerequisite for a successful policy intervention is targeting those high-growth prospect firms that suffer from imperfections in financial markets and therefore need public assistance to realize their full potential. If this is done, the policy measure has a very limited risk of generating a deadweight effect, that is, of crowding out financing from private sources. With respect to the access of young ICT service firms to direct public subsidies, this research investigates whether or not the first prerequisite is satisfied, but nothing can be said here about the second one. Clearly, a failure of policy measures as implemented to distinguish high-growth prospect from low-growth prospect firms on the basis of their genetic characteristics should in itself suggest a reconsideration of Italian technology policy towards young ICT service firms. In fact, channeling public subsidies towards high-growth prospect firms is a necessary, though not sufficient, condition for efficient policy intervention. How can policymakers identify such firms? The extant literature on the growth of high-tech firms has consistently shown that growth performance is related to certain genetic characteristics of high-tech startups. If public subsidies are oriented to these firms, the likelihood of substitution effects clearly decreases. This literature is briefly surveyed in the next section. 2.1. Genetic characteristics Several studies have identified a precise set of characteristics that distinguish high-growth young high-tech firms, and following the approach of Storey (1994) and Almus and Nerlinger (1999), these characteristics can be separated into three distinct groups. The first group consists of founder-specific characteristics. The human capital of a firm’s founders is reputed to be a primary asset for generating competitive advantage in a young technology-intensive startup.4 The available empirical evidence generally lends support to the positive relationship between the competencies of the founders, measured by their level of educational attainment and professional experience, and the growth of their firms (Almus, 2002; Bru¨derl & Preisendo¨rfer, 2000), and this positive relationship is also widely documented in high-tech sectors generally (Almus & Nerlinger, 1999; Cooper & Bruno, 1977; Dahlstrand, 1997; Feeser & Willard, 1990; Nerlinger, 1998; Westhead & Cowling, 1995; see also Colombo & Grilli, 2005b for the Italian case). As for the educational level of the founders, Westhead and Cowling (1995) found that it had a positive and statistically significant impact on firm growth for a sample of new-technology-based firms in the United Kingdom. Colombo and Grilli (2005b) analyzed a large sample of young Italian high-tech enterprises (which includes the sample used in the present study) and provide evidence that graduate education, both in economic-managerial fields and to a lesser extent in technical-scientific fields, has a positive impact on firm growth. However, it is the specific component of the founders’ human capital that exerts the most positive influence on firm performance. Specific human capital is commonly interpreted in the literature (Becker, 1975) as those capabilities of individuals that can directly and profitably be applied to entrepreneurial activities in the newly created firm. It includes industry-specific skills that founders learned in the organization by which they were formerly employed and leadership competencies gained by entrepreneurs, either in a managerial position in another firm or during prior episodes of self-employment (Bru¨derl & Preisendo¨rfer, 2000; Bru¨derl, Preisendo¨rfer, & Ziegler, 1992; Preisendo¨rfer & Voss, 1990). In this respect, Cooper and Bruno (1977) consider young high-tech firms in the San Francisco Peninsula in the 1970s. In a comparison of high growth and discontinued firms, they show that the former were more likely than the latter to have been founded by individuals who came from incubating organizations that operated in the same industry as the new 4

See Cooper and Bruno (1977, p. 21): ‘‘y for a new, high-technology firm, the primary assets are the knowledge and skills of the founders. Any competitive advantage the new firm achieves is likely to be based upon what the founders can do better than others’’.

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

577

firm. Later studies in the 1980s and 1990s generally found similar results for firms operating in low- and medium-tech industries (see for instance Chandler & Jansen, 1992; Cooper, 1985; Dunkelberg, Cooper, Woo, & Dennis, 1987; Siegel, Siegel, & Macmillan, 1993, for exceptions, see Storey, 1994). Furthermore, Feeser and Willard (1990), comparing 39 high-growth firms with a matching set of low-growth firms all operating in the ICT sector, show that the former group is more likely than the latter to have products, markets, and technologies closely related to those of their founders’ incubating organization. Similar results have been reported by Roberts (1992) and Dahlstrand (1997). The former study considers a sample of new high-tech ventures located in the Greater Boston area and shows that a rapid transfer of technology from an advanced technology-based incubating organization and a deep understanding of customer needs and market conditions are key factors in determining a firm’s post-entry performance. The latter study, using a sample of new Swedish high-tech firms, documents that after a short initial period, corporate spinoff firms grow more rapidly than non-spinoff ones. Lastly, Colombo and Grilli (2005b) show that founders’ work experience, in particular when it is gained in the same sector as the startup and in technical departments (R&D, design, engineering, and production), has a positive impact on the growth of young high-tech firms. New technology-based firms have also been found in a number of studies to benefit from the ‘‘leadership experience’’ gained by founders in previous work (see again Colombo and Grilli, 2005b and the literature mentioned there). The last founderspecific characteristic usually considered in the empirical literature is the number of founders. Larger founding teams are reputed to have access to more resources, both tangible (funding) and intangible (competencies); this access allows a firm to grow more rapidly. Generally, the available empirical evidence supports this contention (Colombo & Grilli, 2005b; Eisenhardt & Schoonhoven, 1990; Feeser & Willard, 1990; Teach, Tarpley, & Schwartz, 1986, see Almus & Nerlinger, 1999 for a contrary result).5 The second group of factors that affect the performance of young firms in high-tech sectors consists of firmspecific characteristics. In particular, the ability of a startup to implement and maintain links with external firms and public research institutions has been found in several studies to have a positive effect on its performance (Almus & Nerlinger, 1999; Maskell, 2001; Nerlinger, 1998). Similarly, Chiesa and Piccaluga (2000) and Colombo and Piva (2005) show that new Italian high-tech firms created by people with previous professional experience in academia perform better than the rest of young Italian high-tech enterprises on a series of dimensions, including innovative activity. Similar results have been found to apply in Sweden (Dahlstrand, 1997). Furthermore, if a newly established high-tech firm has received valuable tangible and/or intangible resources from a ‘‘mother’’ company (e.g., complementary technologies, access to distribution channels, after-sale services, support for entry into international markets, or financing), this is likely to have a positive influence on a firm’s performance, especially in its infancy. In the same vein, Colombo, Delmastro, and Grilli (2004) have found that access to these resources positively affects the startup size of the firm. Location in a technology incubator also seems to exert a positive impact on a firm’s growth rate (see Colombo & Delmastro, 2002; Sternberg, 1990; Westhead & Storey, 1994 for Italy, Germany, and the United Kingdom, respectively. For contrary results, see Westhead, 1997 and Westhead & Cowling, 1995 on the topic of new technology-based firms located in the United Kingdom). As for the third group of factors, location-specific variables, the performance of young firms in high-tech sectors is likely to be influenced by the socio-economic characteristics of the area where they are located. More specifically, Colombo and Grilli (2005b) provide evidence that new Italian high-tech ventures benefit from operating in regions with a well-developed infrastructure. Similarly, Audretsch and Dohse (2004) document that location in a region characterized by high availability of knowledge resources in the form of human capital is more conducive to a firm’s growth than location in less well-endowed areas. On the benefits that firms receive from being embedded in innovative high-technology environments, see also Capello (1999) and Breschi and Lissoni (2001). Feldmann (1999) has carried out a survey on the topic. 3. Data set The empirical analysis is based on a sample composed of 351 young Italian ICT service firms. The sample firms were established in 1980 or later (64% were founded since 1995), were independent at the time they were 5

Almus and Nerlinger (1999) do not detect any significant relationship between team foundation and the growth rate of new high-tech firms, even though they do find a positive effect of team foundation in the case of low-and medium-tech firms.

ARTICLE IN PRESS 578

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

founded, were still independent on 1 January 2004 (i.e., they were not controlled by another business organization, even though other organizations may hold minority shares in them), and operate in the following sectors: Internet and telecommunications (TLC) services (Internet service provision, multimedia services, multimedia content, e-commerce, telecommunications services) and software. The sample of young Italian ICT service firms was extracted from the RITA database developed at Politecnico di Milano (Colombo, Delmastro, & Grilli, 2004; Colombo & Grilli, 2005a, b). Data were collected through a series of national surveys administered in the first 6 months of 2000, 2002, and 2004. The sample considered in this paper includes all young ICT service firms from the RITA database that participated in the 2004 survey and for which it was possible to build a complete data set for the variables of interest. The steps in the development of the database are described below. First, Italian firms that matched the above-mentioned criteria for age and sector of operations were identified. For the definition of the target population, a number of sources were used, including lists provided by national industry associations, online and offline directories of commercial firms, and lists of participants in industry trade shows and exhibitions. Information provided by the national financial press, specialized magazines, other sectoral studies, and regional Chambers of Commerce was also considered. Altogether, 1239 ICT service firms were selected for inclusion in the database. For each firm, a contact person (an ownermanager) was also identified. Unfortunately, data provided by official national statistics do not enable a reliable description of the universe of young Italian ICT service enterprises.6 Secondly, a questionnaire was sent to the contact person in each target firm, either by fax or by e-mail. The first section of the questionnaire requested detailed information on the human capital characteristics of the firm’s founders. The second section consisted of further questions about the characteristics of the firm, including its possible access to direct public support measures. Lastly, completed questionnaires were checked for internal coherence by knowledgeable personnel and were compared with published data if the latter were available. In several cases, phone or faceto-face follow-up interviews were conducted with firm owner-managers. This final step was crucial to obtain missing data and to ensure that data were reliable.7 There are no statistically significant differences between the distribution of the sample firms across industries (software, Internet and TLC services) and across geographical areas (northern, central, and southern Italy) and the corresponding distributions of the population of 1239 RITA firms from which the sample was obtained (w2(1) ¼ 0.47 and w2(2) ¼ 4.42, respectively). Note also that the sample is large and heterogeneous, but there is no presumption here that it is random. First, in this domain, representativeness is a slippery notion because new ventures may be defined in various ways (Aldrich, Kallenberg, Marsden, & Cassell, 1989; Birley, 1984; Gimeno, Folta, Cooper, & Woo, 1997). Secondly, as was mentioned earlier, in the absence of reliable official statistics, it is very difficult to identify unambiguously the universe of young Italian ICT service firms. Therefore, one cannot check a posteriori whether or not the sample used in this work is representative of the population. Thirdly, only those firms which survived until the survey date could be included in the sample. In principle, attrition may generate a sample selection bias and distort the estimates. As in most survey-based studies, it is impossible to control fully and properly for this survivorship bias; the best one can do is to check its relevance. To address this concern, Section 5.3.3 is devoted to a statistical analysis that provides a robustness check on the results of the econometric study. 4. Descriptive statistics for access to public subsidies by young ICT service firms The composition of sample firms by sector and geographic area of operation is presented in Table 1. Columns 2 and 3–4 refer to the entire sample of young ICT service firms and the number and percentage of subsidized ones, respectively; columns 5–6 and 7–8 report the number and percentage of firms supported by the central government and by local public entities (e.g., regional governments) respectively. 6

The main problem is that in Italy, most individuals who are defined as ‘‘self-employed’’ by official statistics actually are salaried workers with unconventional employment contracts. Unfortunately, on the basis of official data, such individuals cannot be distinguished from entrepreneurs founding a new firm. 7 Note that there were only two cases in which the set of owner-managers on the survey date did not include at least one of the founders of the firm. For these two firms, information on the human capital characteristics of the founders was checked through interviews with firm personnel to ensure that it did not relate to current owner-managers.

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

579

Table 1 Distribution of subsidized young ICT service firms by industry and geographic area Sector & Area

Firms Subsidized firms

Firms subsidized by central government

Firms subsidized by local administrative entities

N

%

N

%

N

%

N

Sector Internet and TLC services ISP Multimedia services Multimedia content E-commerce TLC services Software

193 28 100 26 27 11 159

68 10 34 15 4 5 61

35.2 35.7 34.0 57.7 14.8 45.5 38.3

36 7 15 8 2 4 44

18.7 25.0 15.0 30.8 7.4 36.3 27.7

45 5 25 8 4 3 37

23.3 17.9 25.0 30.8 14.8 27.3 23.3

Area North Northwest Northeast Center South

235 173 62 57 59

78 57 21 19 32

33.2 33.0 34.4 33.3 54.2

44 34 10 9 27

18.7 19.7 16.1 15.8 45.8

54 39 15 11 17

23.0 22.6 24.2 19.2 28.8

Total

351

129

36.8

80

22.8

82

23.3

Northwest includes the following regions: Liguria, Piemonte, Lombardia, and Val d’Aosta; Northeast: Emilia Romagna, Friuli Venezia Giulia, Trentino Alto Adige, Veneto; Center: Abruzzo, Marche, Lazio, Umbria, Toscana; South: Campania, Basilicata, Molise, Puglia, Calabria, Sicilia, Sardegna.

Overall, a large number of firms (129, or 36.8% of the sample) have benefited from direct public subsidies. The number of firms supported by the central government and by local administrative entities is almost identical (80 and 82, or 22.8% and 23.3% of the sample, respectively). Note also that a non-negligible number of firms have received both forms of direct public support (33, or 9.4% of the sample). Firms operating in the Internet and TLC service sectors have been slightly less subsidized than software houses (35.2% versus 38.3%), with the difference being entirely due to subsidies granted by the central government: 18.7% of Internet and TLC firms gained access to national direct support measures, but for software houses this percentage is 27.7%. Among Internet and TLC service firms, the multimedia content sector shows the greatest inclination to benefit from public assistance (57.7% of the firms are thus supported), both through national (30.8%) and local (30.8%) channels. Firms operating in the TLC services sector show similar percentages: 45.5% of the firms are subsidized, 36.3% and 27.3% at the national and the local level, respectively. The percentages of supported firms in the Internet service provision and multimedia services sectors are very similar (35.7% and 34.0%, respectively), but in the former case, this is mainly due to central government rather than local subsidies (25.0% versus 17.9%), while in the latter case the opposite is true (15.0% versus 25.0%). E-commerce appears to be the least subsidized sector (14.8%), both at the national (7.4%) and the local level (14.8%). Turning to the geographical distribution of sample firms, most of the young ICT service firms are located in northern Italy (235, of which 173 are in the northwest and 62 in the northeast), while central and southern Italy have fewer of these firms (57 and 59, respectively). The northern and central regions of the country have similar features in terms of subsidies for enterprises, while young ICT service firms located in southern Italy show a much greater inclination to benefit from public financial aid than the rest of their Italian competitors (54.2% of the southern firms in the sample have benefited from a direct public subsidy, versus 33.2% and 33.3%, respectively, of the firms located in northern and central Italy). This difference is due mainly to funds granted by the central government (45.8% of southern firms have received this type of subsidy) and can be explained by the large number of laws explicitly intended to promote economic growth in this disadvantaged area.8 Although an assessment of the global efficiency and effectiveness of such mechanisms directed towards 8

The area is defined as ‘‘Mezzogiorno d’Italia’’ and includes the following southern regions: Abruzzo, Molise, Puglia, Campania, Basilicata, Calabria, Sicilia, and Sardegna.

ARTICLE IN PRESS 580

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

Table 2 Distribution of subsidized young ICT service firms by age at the year of first successful subsidy application Firm age (years) in the year of first successful application

Subsidized firms

N 0 1–3 4 and over Totala a

22 44 62 128

% 17.2 34.4 48.4 100

Firms subsidized by central government

Firms subsidized by local administrative entities

N

N

6 22 51 79

% 7.6 27.9 64.5 100

15 31 36 82

% 18.3 37.7 44.0 100

For one firm there is no information on the year of application for the first national subsidy received.

southern Italy is clearly beyond the scope of the paper, in this context it is worth reminding the reader that most young Italian ICT service firms are located in the northern part of the country. As a matter of fact, the likelihood of obtaining access to national public subsidies is much lower for these firms than for their southern counterparts. This tendency is not counterbalanced at all by local public subsidies. Table 2 reports the distribution of subsidized young ICT service firms by age of the firm at the year of application for the first subsidy. Again it is necessary to distinguish between enterprises supported by the central government and by local administrative entities. Overall, 17.2% of the subsidized firms obtained access to public funds at the time they were founded, 34.4% did so at between 1 and 3 years of age, and 48.4% received public support for the first time when they were 4 years old or older. The tendency to support firms at a more mature stage is common to central and local support measures, even if it is greater for the former group. Only 7.6% of the young ICT service firms subsidized by the central government managed to obtain national subsidies in the year they were founded (the corresponding figure is 18.3% for support from local administrative entities), while 64.5% of the sponsored firms obtained national public funds at 4 years of age or older (the corresponding figure is 44.0% for support from local administrative entities). These data, especially those relating to national support schemes, cast serious doubts on the ability of very young firms in the ICT service field to benefit from direct horizontal support mechanisms.9 This is quite worrisome because the seed and startup phases are the most critical phases in a firm’s life. This holds true especially in high-tech sectors, where young and small firms have fewer incentives to invest in R&D due to lack of complementary assets (Teece, 1986) and are the most likely to suffer from imperfections in capital markets (Carpenter & Petersen, 2002). Accordingly, if public support is needed, it is needed especially in the very first years of an ICT service firm’s existence.10 9 As mentioned earlier, the sample does not include firms that failed to survive or to remain independent by the survey date. Therefore there is no information on whether or not these firms received public support in the early stage of their existence. If they did so in relatively greater or smaller numbers than the sampled firms, in principle the results of the econometric analysis might be biased. A robustness check presented in Section 5.3.3 indicates that this risk is very limited. However, as a general remark, a greater tendency of subsidized firms to exit markets would imply either that public support is poisonous, leading to a greater likelihood of failure, or that policymakers are oriented towards supporting extremely weak firms. Alternatively, it might also imply that a firm receiving public support early in its life is less likely to retain its independence. In this latter case, reliance on external seed finance may be likened to an addictive drug: once the drug (funding) is taken, it is necessary to keep providing it, and when governmental authorities are unwilling to provide it, another external source becomes necessary, hence the loss of independence. If this argument can be plausible a priori grounds (and the authors are grateful to W. Edward Steinmueller for the suggestion), the above-mentioned robustness check reveals the opposite tendency, with subsidized firms more likely to retain their independence than nonsubsidized ones. 10 Of course, the question arises why ICT service firms fail to obtain public support in their early years. This interesting issue lies beyond the scope of the present paper. Suffice it here to mention that factors related to both the demand side and the supply side may be at work. On the one hand, due to risk aversion, governmental officials in charge of the delivery of funds may be inclined not to subsidize firms that, because they are very young, experience a relatively higher risk of failure (Lerner, 1999). On the other hand, in order to get public money, top managers in a firm generally need to invest a lot of time and energy (e.g., for information search and administrative procedures and duties), the opportunity cost of which is relatively greater for younger and smaller firms.

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

581

5. Econometric analysis 5.1. Specification of the econometric models To gain insight into the extent to which the genetic characteristics of high-growth prospect firms in hightech industries are also predictors of their access to direct public support programs, a probit model was developed and estimated, in which the dependent variable indicates whether firms have received financial public assistance (1) or not (0) during their existence. The model is of the form: " # k X Pi ¼ F b0 þ bj X ji ¼ F½X 0i b, (1) j¼1

where Pi is the probability of having been subsidized, F is a normal cumulative density function, Xi is the vector of independent variables, and b is the vector of parameters to be estimated. Explanatory variables capture those founder-, firm-, and location-specific characteristics that significantly influence firm performance, plus other control variables (five industry dummies).11 If the risk of substitution effects generated by the horizontal technology policy towards young ICT service firms were low, all the genetic characteristics of high-growth prospect firms should exert a positive influence on the likelihood of obtaining a public subsidy. Then a bivariate probit model was run to investigate whether possible inefficiencies are associated equally with national and local programs. In this case, two equations are defined. In the first equation, the dependent variable indicates whether young ICT service firms have received public financial support through a national program (1) or not (0); in the second equation, the dependent variable indicates whether these firms have received financial support through a local program (1) or not (0). Both equations have the same set of explanatory variables as the probit specification (including the five industry dummies). They are estimated simultaneously, assuming disturbances jointly normally distributed with zero mean and unit variance, and with correlation expressed by the r coefficient (see Greene, 2000 for further details). 5.2. Variables The definition of the explanatory variables is illustrated in Table 3. In accordance with the classification proposed in Section 2, they can be grouped into the following categories. Founder-specific variables. These include the educational attainments of founders; in particular, with regard to university education, a distinction is made between economic and managerial studies (Ecoeduc) and technical and scientific studies (Techeduc).12 Secondly, a distinction is made between founders’ years of work experience in the same sector as the new firm (Specworkexp) and years of work experience in other sectors (Otherworkexp). As mentioned previously, it is widely documented in the empirical literature that it is the specific rather than the generic component of founders’ work experience that positively influences the performance of young high-tech venture firms. The leadership experience possessed by founders is expressed by a proxy variable, DManager. This dummy variable is set to 1 if within the founding team there are one or more individuals who held a managerial position before the establishment of the new firm. Lastly, NFounders is the number of firm founders. In accordance with the argument that the growth prospects of firms are positively affected by the number and the human capital of their founders, and to avoid substitution effects, the above11

The coefficients of these latter are not reported for the sake of simplicity. More precisely, Ecoeduc measures years spent for the attainment of degrees in economics, management, and political sciences, while Techeduc reflects years spent for obtaining degrees in engineering, chemistry, physics, geology, mathematics, biology, medicine, pharmaceutics, and computer science. In order to judge properly the effective level of founders’ competencies, the minimum length of time necessary to attain a certain degree is considered as equivalent to that degree. In order to attain an Italian graduate degree in economics, management, political science, or most scientific fields, 4 years of studies are required, while 5 years is the minimum time for a degree in engineering or chemistry. Master’s and Ph.D. programs require 1 and 3 additional years respectively, independently of the specific field. 12

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

582

Table 3 Determinants of access by young ICT service firms to public support programs Variable

Description

Ecoeduc Techeduc Specworkexp Otherworkexp DManager NFounders DIncubated DMother firm DAcademic

Average number of years of economic and/or managerial education of founders at graduate and postgraduate level Average number of years of scientific and/or technical education of founders at graduate and postgraduate level Average number of years of work experience of founders in the same sector as the startup before founding the firm Average number of years of work experience of founders in other sectors than that of the startup before founding the firm Set to 1 for firms with one or more founders with a prior management position; 0 otherwise Number of founders Set to 1 for firms located in a technology incubator at time of founding; 0 otherwise Set to 1 for firms that at time of founding received some kind of aid from a ‘‘mother’’ firm; 0 otherwise Set to 1 for firms with at least one individual within the founding team with previous work experience in academia or in a public research center; 0 otherwise Value of the index measuring regional infrastructure in 1989 (mean value among Italian regions ¼ 100; source: Confindustria Centro Studi, 1991) Number of years since firm was founded Ratio of the share accounted for by the sector of the new firm out of the total number of high-tech firms that obtained venture capital financing over the period 1997–2003 (source: AIFI) to the share accounted for by the same sector out of the total number of Italian high-tech firms in 2003 (source: RITA Directory) Ratio of the share accounted for by the geographical area in which the new firm is located out of the total number of hightech firms that obtained venture capital financing over the period 1997–2003 (source: AIFI) to the share accounted for by the same geographical area out of the total number of Italian high-tech firms in 2003 (source: RITA Directory)

Locdevelop Age VCSector

VCArea

VCSector and VCArea are defined as follows. As a basis, take the total number of high-tech firms that obtained venture capital financing over the period 1997–2003 (source: AIFI). Let VCSj and VCAk indicate the shares accounted for by sector j and geographical area k out of this number. Let Sj and Ak be the estimated shares accounted for by sector j and geographical area k out of the total number of Italian new technology-based firms in 2003 (source: RITA Directory). Then: VCSectorj ¼ VCSj/Sj and VCAreak ¼ VCAk/Ak.

mentioned founder-specific variables should be positively related to the likelihood of a firm’s obtaining a public subsidy. Firm-specific variables. The dummy variable DIncubated identifies firms located at startup time in technology incubators, while DMother firm is a dummy variable which is set to 1 if a newly established firm received valuable tangible and/or intangible resources at startup time from a ‘‘mother’’ firm (e.g., complementary technologies, access to distribution channels, after-sale services, support for entry into international markets, or financing). DAcademic equals 1 if within the founding team there are one or more individuals with previous work experience as a researcher in academia or in a public research institution. Under the hypothesis of no substitution effects, these variables should exert a positive influence on the likelihood that young ICT service firms can access public subsidies.13 Location-specific variables. Locdevelop reflects the level of economic development in 1989 of the county where each firm is located (Confindustria Centro Studi, 1991). This variable is calculated as the average of the following indices: per capita value added, manufacturing as a percentage of total value added, employment index, per capita bank deposits, automobile-to-population ratio, and consumption of electric power per head. The hypothesis of no substitution effects would entail a positive association between location in areas well supplied with infrastructure and success of new high-tech entrepreneurs in obtaining public subsidies. This obviously does not mean that depressed economic areas may not need more substantial and structural policy interventions aimed at reducing existing gaps in infrastructure and economic performance. Control variables. Age is the difference in years between 2004 and the year a firm was founded. To take into account the availability of private funds to young firms, this group also includes proxy variables representing

13

Note however that policymakers targeting firms with these genetic characteristics may incur a high risk of deadweight effects, because these companies may face less severe obstacles in obtaining external financing from private sources.

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

583

Table 4 Descriptive statistics for the explanatory variables in the econometric models Variable

Mean

S.D.

Min

Max

Ecoeduc Techeduc Specworkexp Otherworkexp DManager NFounders DIncubated DMother firm DAcademic Locdevelop Age VCSector VCArea

0.411 1.718 3.314 7.146 0.273 2.864 0.077 0.059 0.119 114.244 8.585 1.396 1.027

1.030 2.139 5.356 7.457 0.446 1.683 0.267 0.237 0.324 28.321 5.483 1.525 1.141

0.000 0.000 0.000 0.000 0.000 1.000 0.000 0.000 0.000 43.700 0.000 0.000 0.000

5.000 8.000 27.000 49.000 1.000 15.000 1.000 1.000 1.000 174.700 23.000 16.222 5.892

the propensity of the venture capital industry to invest in the sector (VCSector) and the geographical area (VCArea) in which the new firm operates.14 Table 4 presents descriptive statistics for the explanatory variables, and Table 5 their correlation matrix. 5.3. Results 5.3.1. Probit model results Probit model results are shown in Table 6. Overall, the statistical significance of the model is very low (McFadden R2 is equal to 0.093), and with the sole exception of location at startup time in a technology incubator, all the founder- and firm-specific attributes that characterize young successful enterprises in hightech sectors show very poor explanatory power for the success of young Italian ICT service firms in accessing public funds. The number of founders, the level of founders’ specific and generic work experience, and the level of entrepreneurs’ ‘‘leadership experience’’ gained in previous managerial employment do not perform any role in discriminating subsidized from non-subsidized enterprises. As for firm-specific characteristics, the only statistically significant coefficient (at the 95% confidence level) turns out to be that of DIncubated, pointing to the role of ‘‘bridging institution’’ between the public sector and the entrepreneurial world that technology incubators may play (Colombo & Delmastro, 2002). In fact, technology incubators, especially in areas characterized by a weak innovation system, may ease the information flow from legislators to entrepreneurs and may offer consultancy services to incubated firms, helping the latter fill out application forms and prepare well-designed proposal submissions. This result may also suggest that these institutions can exert a signaling function and a certification effect with respect to policymakers. The other firm-specific variables, DMother firm and DAcademic, are not significant. As for the location-specific variable, confirming the substantial advantage that young ICT service enterprises in southern Italy have in accessing public funds, the coefficient of Locdevelop, which captures the level of infrastructure development in the county where the firm is located, is negative and statistically significant at the 99% confidence level. In other words, direct public aid flows more intensively towards young ICT service firms located in less-developed areas, where the empirical literature documents that location in an area less well equipped in terms of infrastructure and knowledge resources may depress a firm’s growth rate. Overall, the picture that emerges is one of random allocation of public funds to young ICT service firms, with no relationship between the genetic characteristics that signal high-growth prospect firms and the likelihood of 14

VCSector and VCArea are defined as follows. The starting point was the total number of Italian high-tech firms that obtained venture capital financing between 1997 and 2003 (source: AIFI). Let VCSj and VCAk represent the shares of this number accounted for by sector j and geographical area k. Let Sj and Ak be the estimated shares of sector j and geographical area k of the total number of Italian newtechnology-based firms in 2003 (source: RITA Directory). Then VCSectorj ¼ VCSj/Sj and VCAreak ¼ VCAk/Ak.

584

Variable

Ecoeduc Techeduc Specworkexp Otherworkexp DManager NFounders DIncubated DMother firm DAcademic Locdevelop Age

1.00 0.02 0.06 0.09 0.10 0.02 0.09 0.33 0.07 0.02 0.08 0.03

1.00 0.38 0.16 0.03 0.06 0.12 0.21 0.00 0.03 0.02 0.07

1.00 0.11 0.03 0.02 0.03 0.02 0.07 0.03 0.12 0.05

1.00 0.15 0.00 0.18 0.08 0.10 0.07 0.14 0.03

1.00 0.10 0.05 0.27 0.01 0.03 0.03 0.13

1.00 0.07 0.20 0.17 0.17 0.08 0.28

1.00 0.01 0.01 0.08 0.03 0.01

1.00 0.12 0.15 0.14 0.07

1.00 0.03 0.12 0.32

1.00 0.31 1.00 0.06 0.04

1.00

ARTICLE IN PRESS

Ecoeduc 1.00 Techeduc 0.12 Specworkexp 0.07 Otherworkexp 0.01 DManager 0.11 NFounders 0.00 DIncubated 0.07 DMother firm 0.04 DAcademic 0.02 Locdevelop 0.13 Age 0.12 PESector 0.06 PEArea 0.07

VCSector VCArea

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

Table 5 Correlation matrix of the explanatory variables in the econometric models

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

585

Table 6 Probit model of the access of young ICT service firms to public financial support programs Model

a0 a1 a2 a3 a4 a5 a6 a7 a8 a9 a10 a11 a12 a13

Constant Ecoeduc Techeduc Specworkexp Otherworkexp DManager NFounders DIncubated DMother firm DAcademic Locdevelop Age VCSector VCArea No. of observations Log-likelihood R2 McFadden

Coefficient

Standard error

0.034 0.119 0.058 0.021 0.012 0.109 0.062 0.602** 0.288 0.101 0.008*** 0.047*** 0.044 0.030

0.437 0.074 0.038 0.016 0.011 0.181 0.045 0.291 0.320 0.262 0.003 0.015 0.038 0.069 337 201.976 0.093

*po0.10; **po0.05; ***po0.01. For the sake of simplicity, estimated coefficients for industry dummy variables are not reported. For 14 firms, information on the variables of interest was not available.

obtaining public subsidies. Almost all the variables that influence firm performance are found to have very poor explanatory power for a company’s capacity to access public direct support programs. The risk is, therefore, that such policy measures have generated conspicuous substitution effects in the ICT services market. 5.3.2. Bivariate probit model results To gain further insight into this issue, and in particular to verify whether there are perceptible differences between national and local programs, a bivariate probit model was estimated, and the results are reported in Table 7. The low explanatory power of founder- and firm-specific characteristics is strongly confirmed, suggesting that inefficiencies are associated with both sources of public funds. Note also that the correlation coefficient r between the error terms of the two equations is positive and statistically significant at the 99% confidence level, suggesting that access by young ICT service firms to national and to local public subsidies are interdependent events. In other words, some unobserved characteristic of the firms (unrelated to the individual genetic characteristics of high-growth prospect firms) favors access to both types of subsidy. Turning to the analysis of the impact of the explanatory variables, in this case DIncubated has a positive and statistically significant influence on access to national programs only, but not on access to local programs. A firm’s location in a technology incubator may help in accessing national subsidies that may have a more complex and competitive selection procedure than local ones. Furthermore, the allocation of central government funding is positively influenced by the technical education of the firm’s founders, while the presence within the founding team of an entrepreneur with previous managerial experience has a positive impact on the probability of a firm’s access to funds provided by a local administrative entity. All the other founder- and firm-specific covariates do not have any impact on the likelihood of benefiting from either national or local direct support mechanisms, except for Specworkexp that shows a negative and statistically significant impact (at the 90% level) on the probability of obtaining local subsidies. As for location-specific characteristics, Locdevelop presents a negative and statistically significant coefficient for central government programs and a negative but not statistically significant coefficient for local ones. The result may be due to the high number of national industrial subsidy programs which offer exclusive or preferential access to firms located in southern Italy in order to reduce the gap between this region and the more industrialized areas of central and especially northern Italy. If clearly support for economic depressed areas is an industrial policy priority, firm-level

ARTICLE IN PRESS 586

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

Table 7 Bivariate probit model of the access of young ICT service firms to public support measures: national vs. local programs Model Central funding

a0 a1 a2 a3 a4 a5 a6 a7 a8 a9 a10 a11 a12 a13

Constant Ecoeduc Techeduc Specworkexp Otherworkexp DManager NFounders DIncubated DMother firm DAcademic Locdevelop Age VCSector VCArea r No. of observations Log-likelihood

Local funding

Coefficient

Standard error

0.232 0.176 0.128*** 0.004 0.008 0.171 0.057 0.699** 0.061 0.184 0.012*** 0.054*** 0.061 0.036

0.511 0.107 0.046 0.023 0.013 0.243 0.056 0.353 0.495 0.330 0.003 0.018 0.047 0.087

Coefficient

0.847* 0.074 0.009 0.034* 0.019 0.504** 0.065 0.310 0.024 0.224 0.003 0.037** 0.009 0.007 0.403 (0.106)*** 337 324.622

Standard error 0.498 0.081 0.047 0.020 0.013 0.202 0.048 0.362 0.431 0.283 0.003 0.017 0.045 0.079

*po0.10; **po0.05; ***po0.01. For the sake of simplicity, estimated coefficients for industry dummy variables are not reported. For 14 firms, information on the variables of interest was not available.

subsidies may be of little relevance unless these measures form part of a better-articulated policy approach with structural interventions, especially on the infrastructure side. As for control variables, Age has a positive and statistically significant coefficient in both equations.15 5.3.3. Robustness testing Finally, as mentioned in Section 3, the data do not enable proper control for possible selectivity bias generated by firm failures and loss of independence. What can be reasonably done is to check the likely extent of this bias indirectly. For this purpose, attention was focused on the sample of young ICT service firms from the RITA 2000 database. This sample contains 263 young ICT service firms (see Colombo et al., 2004). In this sample, 69 firms had received one or more public subsidies from the date they were founded until the end of 1999. The authors examined the exit rate of these 263 firms from 2000 to 2003 due either to bankruptcy or to mergers and acquisitions. 11 subsidized young ICT service firms, or 15.9% of the subsidized sub-sample, exited the market during this period. The corresponding percentage for firms that obtained no subsidies was substantially higher (32.4%, or 63 firms). A w2 test shows that there are statistically significant differences at conventional confidence levels between the two sub-samples (w2(1) ¼ 6.87). Actually, the difference between the two percentages was almost entirely attributable to firms that lost their independence. Looking at the firms that went bankrupt, the percentage of bankruptcies among the subsidized firms was 10.1%, while the percentage for non-subsidized ones was 15.4%. In this case, a w2 test shows that there is no statistically significant difference at conventional confidence levels between the two sub-samples (w2(1) ¼ 1.19). The corresponding percentages for young ICT service firms that were merged or acquired were 5.8% and 17.0%, respectively for subsidized and nonsubsidized firms, with the difference being significant at the 95% confidence level (w2(1) ¼ 5.29). In other words, the difference for firms that went bankrupt was not statistically significant. On the other hand, firms that obtained public subsidies were far more likely to retain their independence than other firms. This means that the sample 15 This result suggests that neither national nor local subsidies are particularly effective in targeting very young ICT service firms. In fact, if public subsidies were provided mainly to firms in the very early stage of their existence, the coefficient of the age variable would be not statistically different from zero. This does not occur in this case.

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

587

used in this study might over represent subsidized firms because of their greater capacity to remain independent. To analyze the possible implications of these results on the estimates, a comparison was made between the two groups of exited young ICT service firms (subsidized versus nonsubsidized) to investigate whether they differed with respect to the explanatory variables used in the models. The results of t-tests and w2 tests run on the variables reflecting genetic high-growth firm characteristics revealed no statistically significant differences (except for the number of founders at p-valueo0.10) between the two groups (see Appendix A1). Thus, while it is fair to acknowledge that the sample used in this study suffers from a survivorship bias, it can be confidently asserted that this does not greatly influence the results. 6. Concluding remarks 6.1. Summary of results This research intended to investigate whether horizontal programs with general objectives (e.g., a policy targeting support to innovation or entrepreneurship) may be efficient in supporting young ICT service firms, or conversely, whether this type of enterprise needs more specific, customized programs (i.e., a vertical technology policy approach). Young high-tech firms deserve particular attention, given their strategic role in overall economic growth and the ‘‘near-universal recognition of the presence of market failures in the provision of finance for new technology-based firms’’ (Storey & Tether, 1998, p. 1049; see also Carpenter & Petersen, 2002). Public intervention should aim to sustain high-tech entrepreneurship while limiting the risk of market distortions. This means targeting those enterprises that possess genetic characteristics of high-growth potential but may experience difficulties in realizing their business objectives because of lack of financial resources. By taking this approach, policy measures avoid the risk of generating substitution or deadweight effects in markets. The former effect arises when inefficient but subsidized firms manage to stay in the market and this incumbency crowds out (in terms of survival and growth performance) potentially efficient but nonsubsidized firms. The second effect arises when beneficiaries do not need subsidies and subsidy financing simply crowds out other private funds. For the purpose of the paper, the Italian experience is very interesting. In fact, in Italy, there has never been a direct support program exclusively targeted to young high-tech firms; moreover, there has never been systematic indirect support for this type of firm. Certain questions then arise: to what extent are young ICT service firms able to obtain funds from untargeted programs? And more importantly, is public money efficiently allocated to young ICT service firms through direct horizontal programs? This research took advantage of a new data set covering a sample of 351 young Italian firms operating in the ICT service sector (Internet service provision, multimedia services, multimedia content generation, e-commerce, telecommunication services, and software). Firms in the sample were established in 1980 or later, were independent at the time they were founded, and were still independent on 1 January 2004 (i.e., they were not controlled by another business organization, even though other organizations may hold minority shares). The main results of this analysis are summarized below. First, despite the untargeted nature of most policy programs, the number of young ICT service firms receiving direct support is considerable (36.8%). The percentages of firms supported at the national and local levels are 22.8% and 23.3%, respectively, 9.4% of the firms have benefited from programs at both levels. Secondly, general-purpose programs appear to be far from efficient in offering direct support to young ICT service firms. While most Italian firms in the ICT service sector are located in the northern part of the country, enterprises in the south are much more likely to obtain access to policy support. This is a consequence of the large number of laws explicitly intended to support this disadvantaged area. In addition, almost half of subsidized young ICT service firms benefited from public funding when they were 4 years old or older at the time of application. This is particularly true for national programs, and it is worthy of attention by policymakers since high-tech firms are likely to suffer more severely from market imperfections when they are young and lack a track record. Finally, the econometric analysis highlighted that, with very few exceptions, none of the founderand firm-specific characteristics associated with successful young ICT service firms are determinants of access by these firms to direct policy support programs, whether at the national or local level. This observation points to the high risk that a technology policy based predominantly on horizontal measures, such as the Italian policy, might generate important substitution effects in the ICT service sector.

ARTICLE IN PRESS 588

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

6.2. Policy implications To sum up, the findings of this research support the view that a direct horizontal approach may be insufficient and inefficient in supporting young ICT service firms. This argument seems to hold (at least in the Italian context) irrespective of the local or central nature of the public institution granting the subsidy. In other words, the geographical dimension of a policy measure does not play a relevant role in the efficiency of the interventions directed at young ICT service firms as long as it does not provide specific and customized aid to this type of firm. Given the global market in which these firms are called upon to operate and the national priority accorded to the development of efficient young ICT service firms in the whole country, the duplication of efforts at national and local levels may be unnecessary. As a corollary, policymakers would be better advised to focus on vertical measures at the national level that take explicit account of the specific characteristics of this category of firms and of high-tech startups generally. In particular, in this field, government officials are very likely to face severe difficulties in selecting and monitoring the beneficiaries of public support among young high-tech enterprises. Accordingly, indirect support programs that delegate this function to specialized institutions are likely to be more efficient. Two types of indirect measures are worthy of attention in the authors’ view. First, in countries like Italy that are lagging behind in high-tech activities, policymakers should encourage the development of effective technology incubators. In fact, in these countries there are substantial market failures regarding the provision of essential inputs to young high-tech enterprises (e.g., real estate and technical and other business services, in addition to financing). Technology incubators reduce the transaction costs associated with the purchase of these inputs; in addition, while screening candidates for location in the incubator, they provide an evaluation of their quality. In so doing, they perform an important bridging function between new high-tech firms and the financial community, which is much more valuable in these cases than in countries where the national innovation system is more advanced (Colombo & Delmastro, 2002). Secondly, in countries with a bank-based financial system, the supply of venture capital to new high-tech enterprises, especially in the seed and startup stages, is limited. Development of an efficient venture capital industry should therefore be a key target of technology policy. In this perspective, public co-financing of private equity investments would be helpful. Nonetheless, one of the reasons that deter the provision of venture capital to small young high-tech firms is the high fixed cost of screening investment proposals and monitoring the operations of the firms that are financed. Therefore, measures aimed at reducing the managerial and administrative costs of venture capitalists so as to make investments in these firms more attractive would also be important. In this respect, the guidelines provided in the European industrial policy by the Commission of the European Communities (2002) offered the possibility for the implementation of more specific and targeted policy measures than those allowed in the past (see Sterlacchini, 2004 for a discussion). In Italy, the Piano per l’Innovazione Digitale nelle Imprese (‘‘Plan for digital innovation in the industrial sector’’), launched in 2003 by the Ministro per l’Innovazione e le Tecnologie and Ministro delle Attivita` Produttive, which aims to provide indirect support to selected small enterprises in the ICT sector and also to facilitate investments by venture capitalists, is a move in the right direction. The recommendation here is to improve, coordinate, and consolidate the implementation of this type of policy measure. Acknowledgments The authors wish to thank Ed Steinmueller, Robin Mansell, and participants in the EURO CPR 2005 Conference for helpful comments. Responsibility for any errors lies solely with the authors. The authors are jointly responsible for the work. However, Sections 1 and 6 were written by Massimo G. Colombo and the remaining sections by Luca Grilli. Appendix A1 A robustness check: a comparison of genetic high-growth firm characteristics between subsidized and non-subsidized exited young ICT service firms (Table A1).

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

589

Table A1

Ecoeduc Techeduc Specworkexp Otherworkexp DManager NFounders DIncubated DMother firm DAcademic Locdevelop

Exited subsidized

Exited non-subsidized

Mean difference test

0.454 0.934 2.545 6.987 18.1% 4.091 18.1% 9.1% 0.0% 111.282

0.748 1.024 5.251 9.313 34.9% 2.269 7.9% 7.9% 4.8% 122.303

0.964 0.207 1.383 1.250 1.197 1.965* 1.148 0.017 0.546 1.017

Mean difference tests refer to t-test for continuous variables and w2-test for percentages. *po0.10; **po0.05; ***po0.01.

References Acs, Z. J. (2004). The value of entrepreneurial start-ups to an economy. Seminar Discussion Paper, Diebold Institute for Public Policy Studies. Aldrich, H. E., Kallenberg, A., Marsden, P., & Cassell, J. (1989). In pursuit of evidence: Sampling procedures for locating new businesses. Journal of Business Venturing, 4, 367–386. Almus, M. (2002). What characterizes a fast growing firm? Applied Economics, 34, 1497–1508. Almus, M., & Nerlinger, E. A. (1999). Growth of new technology-based firms: Which factors matter? Small Business Economics, 13, 141–154. Asquith, P., & Mullins, D. W. (1986). Equity issues and offering dilution. Journal of Financial Economics, 15, 61–89. A˚stebro, T., & Bernhardt, I. (1999). The winner’s curse of human capital. Working Paper CES 99-5, Center for Economic Studies, US Department of Commerce. Audretsch, D., & Dohse, D. (2004). The impact of location on firm growth. CEPR Discussion Paper no. 4332. Becker, G. S. (1975). Human capital. New York: National Bureau of Economic Research. Birley, S. (1984). Finding the new firm. Proceedings of the Academy of Management Meetings, 47, 64–68. Black, J., de Meza, D., & Jeffreys, D. (1996). House pricing, the supply of collateral and the enterprise economy. Economic Journal, 106, 60–75. Blanchflower, D., & Oswald, A. (1998). What makes an entrepreneur. Journal of Labor Economics, 16, 26–60. Breschi, S., & Lissoni, F. (2001). Knowledge spillovers and local innovation systems: A critical survey. Industrial and Corporate Change, 10, 975–1005. Bru¨derl, J., & Preisendo¨rfer, P. (2000). Fast growing businesses: Empirical evidence from a German study. International Journal of Sociology, 30, 45–70. Bru¨derl, J., Preisendo¨rfer, P., & Ziegler, R. (1992). Survival chances of newly founded business organizations. American Sociological Review, 72, 227–242. Capello, R. (1999). Spatial transfer of knowledge in high-technology milieux: Learning vs. collective learning processes. Regional Studies, 33, 353–365. Carpenter, R. E., & Petersen, B. C. (2002). Capital market imperfections, high-tech investment, and new equity financing. Economic Journal, 112, F54–F72. Chandler, G. N., & Jansen, E. (1992). The founder’s self-assessed competence and venture performance. Journal of Business Venturing, 7, 223–236. Colombo, M. G., & Delmastro, M. (2002). How effective are technology incubators? Evidence from Italy. Research Policy, 31, 1103–1122. Colombo, M. G., Delmastro, M., & Grilli, L. (2004). Entrepreneurs’ human capital and the start-up size of new technology-based firms. International Journal of Industrial Organization, 22, 1183–1211. Colombo, M.G., & Grilli, L. (2005a). Start-up size: The role of external financing. Economics Letters, 88, 243–250. Erratum (2006), 90, 148. Colombo, M. G., & Grilli, L. (2005b). Founders’ human capital and the growth of new technology-based firms: A competence-based view. Research Policy, 34, 795–816. Colombo, M. G., & Grilli, L. (2007). Funding gaps? Access to bank loans by high-tech start-ups. Small Business Economics, 29, 25–46. Colombo, M. G., & Piva, E. (2005). Academic start-ups and new technology-based firms: A matched pair comparison. In Paper presented at fifth triple helix conference, 18–21 May 2005, Turin, Italy. Chiesa, V., & Piccaluga, A. (2000). Exploitation and diffusion of public research: The case of academic spin-off firms in Italy. R&D Management, 30, 329–339. Commission of the European Communities (2002). Industrial Policy in an Enlarged Europe. Bruxelles, Communication from the Commission, COM (2002) 714—12 December 2002. Confindustria Centro Studi. (1991). Indicatori economici provinciali. SIPI, Roma: Collana Industria e Territorio.

ARTICLE IN PRESS 590

M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

Cooper, A. C. (1985). The role of incubator organizations in the founding of growth-oriented firms. Journal of Business Venturing, 1, 75–86. Cooper, A. C., & Bruno, A. V. (1977). Success among high-technology firms. Business Horizons, 20, 16–22. Dahlstrand, A. L. (1997). Growth and inventiveness in technology-based spin-off firms. Research Policy, 26, 331–344. Daveri, F., & Silva, O. (2004). Not only Nokia: What Finland can tell us about new economy growth. Economic Policy, 38, 117–163. Dunkelberg, W. G., Cooper, A. C., Woo, C., & Dennis, W. J. (1987). New firm growth and performance. In N. C. Churchill, J. A. Hornaday, B. A. Kirchhoff, C. J. Krasner, & K. H. Vesper (Eds.), Frontiers of entrepreneurship research. Boston, MA: Babson College. Eisenhardt, K. M., & Schoonhoven, C. B. (1990). Organizational growth: Linking founding team, strategy, environment and growth among US semiconductor ventures, 1978–1988. Administrative Science Quarterly, 35, 504–529. European Commission Joint Research Centre (2002). RTD Evaluation Toolbox, Socio-Economic Evaluation of Public RTD Policies. Evans, D., & Jovanovic, B. (1989). An estimated model of entrepreneurial choice under liquidity constraints. Journal of Political Economy, 97, 808–827. Evans, D., & Leighton, L. (1989). Some empirical aspects of entrepreneurship. American Economic Review, 79, 519–535. Feeser, H. R., & Willard, G. E. (1990). Founding strategy and performance: A comparison of high and low growth high tech firms. Strategic Management Journal, 11, 87–98. Feldmann, M. P. (1999). The new economics of innovation, spillovers and agglomeration: A review of empirical studies. Economics of Innovation and New Technology, 8, 5–26. Gimeno, J., Folta, T., Cooper, A., & Woo, C. (1997). Survival of the fittest? Entrepreneurial human capital and the persistence of under performing firms. Administrative Science Quarterly, 42, 750–783. Gordon, R. J. (2000). Does the new economy measure up to the great inventions of the past? Journal of Economic Perspective, 14, 49–74. Greene, W. H. (2000). Econometric analysis (4th ed). Englewood Cliffs, NJ: Prentice-Hall. Griliches, Z. (1992). The search for R&D spillovers. Scandinavian Journal of Economics, 94(suppl.), S29–S47. Grilli, L. (2005). Internet start-ups access to the bank loan market: Evidence from Italy. Applied Economics, 37(3), 293–305. Guillen, M. F., & Suarez, S. L. (2001). Developing the Internet: Entrepreneurship and public policy in Ireland, Singapore, Argentina, and Spain. Telecommunications Policy, 25, 349–371. Holtz-Eakin, D. (2000). Public policy toward entrepreneurship. Small Business Economics, 15, 283–291. Holtz-Eakin, D., Joulfaian, D., & Rosen, H. S. (1994a). Entrepreneurial decisions and liquidity constraints. Rand Journal of Economics, 25, 334–347. Holtz-Eakin, D., Joulfaian, D., & Rosen, H. S. (1994b). Sticking it out: Entrepreneurial survival and liquidity constraints. Journal of Political Economy, 102, 53–75. Jaffe, A.B. (1996). Economic analysis for research spillovers: Implications for the advance technology program. In Paper of the National Institute of Standards and Technology, Advanced Technology Program. Washington DC: US Department of Commerce. Lee, I., Lochhead, S., Ritter, J., & Zhao, Q. (1996). The costs of raising capital. Journal of Financial Research, 19, 59–74. Lerner, J. (1999). The Government as venture capitalist: The long-run impact of the SBIR Program. Journal of Business, 72, 285–318. Lerner, J. (2002). When bureaucrats meet entrepreneurs: The design of effective ‘public venture capital’ programmes. Economic Journal, 112, F73–F84. Lerner, J., & Kegler, C. (2001). Evaluating the small business innovation research program: A literature review. In C. Wessner (Ed.), The small business innovation research program (SBIR): An assessment of the department of defense fast track initiative (pp. 397–426). Washington DC: National Academy Press. Maskell, P. (2001). Towards a knowledge-based theory of the geographical cluster. Industrial and Corporate Change, 10, 921–943. Meyer, B.D. (1990). Why are there so few black entrepreneurs? NBER Working Paper 3537, Cambridge, MA. Myers, S., & Majluf, N. (1984). Corporate financing and investment decisions when firms have information that investors do not have. Journal of Financial Economics, 13, 197–221. Nerlinger, E. (1998). Standorte und entwicklung junger innovativer unternehmen: Empirische ergebnisse fuer West-Deutschland, mimeo. Organization for Economic Co-Operation and Development (OECD) (2002). Measuring the Information Economy. Paris. Organization for Economic Co-Operation and Development (OECD) (2004). The Economic Impact of ICT-measurement, Evidence and Implications. Paris. Preisendo¨rfer, P., & Voss, T. (1990). Organizational mortality of small firms: The effects of entrepreneurial age and human capital. Organizational Studies, 11, 107–129. Roberts, E. B. (1992). The success of high-technology firms: Early technological and marketing influences. Interfaces, 22, 3–12. Santarelli, E., & Vivarelli, M. (2002). Is subsidizing entry an optimal policy? Industrial and Corporate Change, 11, 39–52. Schreyer, P. (2000). The contribution of information and communication technology to output growth: A study of the G7 Countries. OECD STI Working Paper, 2000/2, OECD, Paris. Siegel, D., Wessner, C., Binks, M., & Lockett, A. (2003). Policies promoting innovation in small firms: Evidence from the US and UK. Small Business Economics, 20, 121–127. Siegel, R., Siegel, E., & Macmillan, I. C. (1993). Characteristics distinguishing high growth ventures. Journal of Business Venturing, 8, 169–180. Sterlacchini, A. (2004). Ricerca ed alta tecnologia in Italia: Le basi per un rilancio. L’Industria, 25, 675–705. Sternberg, R. (1990). The impact of innovation centers on small technology-based firms: The example of the Federal Republic of Germany. Small Business Economics, 2, 105–118. Storey, D. J. (1994). Understanding the small business sector. London: Thomson Learning.

ARTICLE IN PRESS M.G. Colombo, L. Grilli / Telecommunications Policy 31 (2007) 573–591

591

Storey, D. J., & Tether, B. S. (1998). Public policy measures to support new technology-based firms in the European Union. Research Policy, 26, 1037–1057. Teach, R. D., Tarpley, F. A., & Schwartz, R. G. (1986). Software venture teams. In R. Ronstadt, J. Hornaday, R. Peterson, & K. H. Vesper (Eds.), Frontiers in entrepreneurial research (pp. 546–562). Wellesley, MA: Babson Center for Entrepreneurial Studies. Teece, D. J. (1986). Profiting from technological innovation: Implications for integration, collaboration, licensing and public policy. Research Policy, 15, 285–305. Westhead, P. (1997). R&D ‘‘inputs’’ and ‘‘outputs’’ of technology-based firms located on and off science parks. R&D Management, 21, 109–123. Westhead, P., & Cowling, M. (1995). Employment change in independent owner-managed high technology firms in Great Britain. Small Business Economics, 7, 111–140. Westhead, P., Storey, D. J. (1994). An assessment of firms located on- and off-science parks in the UK. HMSO Paper, London. Westhead, P., & Storey, D. J. (1997). Financial constraints on the growth of high technology small firms in the United Kingdom. Applied Financial Economics, 7, 197–201.